{"id":105450,"date":"2022-11-03T09:36:47","date_gmt":"2022-11-03T07:36:47","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/ispolzovanie-mashinnogo-obucheniya-dlya-opredeleniya-emoczij-i-kontrolya-nad-svoej-mimikoj"},"modified":"2022-11-03T09:36:47","modified_gmt":"2022-11-03T07:36:47","slug":"ispolzovanie-mashinnogo-obucheniya-dlya-opredeleniya-emoczij-i-kontrolya-nad-svoej-mimikoj","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/news\/ispolzovanie-mashinnogo-obucheniya-dlya-opredeleniya-emoczij-i-kontrolya-nad-svoej-mimikoj","title":{"rendered":"P\u00ebrdorimi i m\u00ebsimit t\u00eb makinerive p\u00ebr t\u00eb identifikuar emocione dhe p\u00ebr t\u00eb kontrolluar shprehjen e fytyr\u00ebs","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Andrej Sav\u010denko nga filiali i Nizhni Novgorod i Shkoll\u00ebs s\u00eb Lart\u00eb t\u00eb Ekonomis\u00eb publikoi rezultatet e hulumtimeve t\u00eb tij n\u00eb fush\u00ebn e m\u00ebsimit t\u00eb makinave, t\u00eb lidhura me njohjen e emocioneve n\u00eb fytyrat e njer\u00ebzve, t\u00eb pranish\u00ebm n\u00eb fotografi dhe video. Kodi \u00ebsht\u00eb shkruar n\u00eb gjuh\u00ebn Python duke p\u00ebrdorur PyTorch dhe shp\u00ebrndahet n\u00ebn licenc\u00ebn Apache 2.0. Disponohen disa modele t\u00eb gatshme, p\u00ebrfshir\u00eb ato t\u00eb p\u00ebrshtatshme p\u00ebr p\u00ebrdorim n\u00eb pajisje mobile. <\/p>\n
N\u00eb baz\u00eb t\u00eb bibliotek\u00ebs, nj\u00eb zhvillues tjet\u00ebr krijoi programin sevimon, i cili lejon monitorimin e ndryshimeve t\u00eb emocioneve duke p\u00ebrdorur nj\u00eb kamer\u00eb video dhe ndihmon n\u00eb kontrollin e tensionit t\u00eb muskujve t\u00eb fytyr\u00ebs, p\u00ebr shembuj, p\u00ebr t\u00eb eliminuar mbip\u00ebrshkarkimin, ndikimin e drejtp\u00ebrdrejt\u00eb n\u00eb humor dhe, n\u00eb p\u00ebrdorim t\u00eb gjat\u00eb, parandalimin e shfaqjes s\u00eb rrudhave mimike. P\u00ebr t\u00eb p\u00ebrcaktuar pozicionin e fytyr\u00ebs n\u00eb video, \u00ebsht\u00eb p\u00ebrdorur biblioteka CenterFace. Kodi sevimon \u00ebsht\u00eb shkruar n\u00eb Python dhe shp\u00ebrndahet n\u00ebn licenc\u00ebn AGPLv3. N\u00eb ngarkimin e par\u00eb, modelet shkarkohen, pas s\u00eb cil\u00ebs programi nuk k\u00ebrkon lidhje me internetin dhe funksionon plot\u00ebsisht autonom. Jan\u00eb p\u00ebrgatitur udh\u00ebzime p\u00ebr nisjen n\u00eb Linux\/UNIX dhe Windows, si dhe nj\u00eb imazh docker p\u00ebr Linux. <\/p>\n
Sevimon funksionon n\u00eb k\u00ebt\u00eb m\u00ebnyr\u00eb: fillimisht n\u00eb imazhin nga kamera identifikohet fytyra, pastaj fytyra krahasohet me \u00e7do nga tet\u00eb emocionet (merzi, p\u00ebrbuzje, neveri, frik\u00eb, g\u00ebzim, munges\u00eb emocionesh, trishtim, befasim), pas s\u00eb cil\u00ebs p\u00ebr secil\u00ebn emocion jepet nj\u00eb vler\u00ebsim i ngjashm\u00ebris\u00eb. Vlerat e marra ruhen n\u00eb nj\u00eb ditar n\u00eb format tekstual p\u00ebr analiz\u00ebn e m\u00ebtejshme nga programmi sevistat. P\u00ebr secil\u00ebn emocion n\u00eb skedarin e konfigurimit mund t\u00eb vendosen kufij t\u00eb sip\u00ebrm dhe t\u00eb posht\u00ebm, ku kalimi i t\u00eb cil\u00ebve jep nj\u00eb kujtes\u00eb menj\u00ebher\u00eb. \n Burimi: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"